OpenAI releases 722 AI-generated math manuscripts with partial Lean proofs
OpenAI · x · 2026-10-07
OpenAI announced a broad release of new mathematical results from an internal frontier model, published on the openai/math GitHub repo.
Key points:
- The collection contains 722 manuscripts with supporting proof artifacts (preprints, reasoning traces, and a growing set of Lean formalizations), under Apache-2.0.
- Results are at varying verification stages: not all have Lean formalizations, and OpenAI acknowledges unformalized results may contain issues that it will fix over time.
- Context: OpenAI expanded to open research problems after its existing math evals saturated; some outputs build on earlier model-generated results.
- The release strategy was shaped by advice from the independent Advisory Group on Mathematics and AI at the Institute for Advanced Study.
More from Companies & People
- Reve is hiring a product lead to rethink HCI, on-site in Palo Alto — cantrell · 2026-10-07
- There's No Such Thing as an AI-Ready Culture, HBR Argues — rseroter · 2026-10-07
- How to hire in the age of AI: 7 experiments on evaluating talent with capable agents — vasuman · 2026-10-07
- From intern to engineer: Manat demos Tandem Voice at keynote — ZackRW · 2026-10-07
- Bun's Rust rewrite once drew weeks of pushback; now 5+ TypeScript toolchain rewrites are normal — cnakazawa · 2026-10-07
- FICO to cut ~15% of workforce as it embeds AI into product development — Polymarket · 2026-10-07